Position detection for arbitrary-oriented ships in satellite imagery via convolutional neural network

Combining satellite images with computer vision techniques, has enabled us to develop a plethora of applications extending from precise agriculture to urban planning. In this project, ship detection from satellite images using artificial neural networks is focused on. This is an important task which can entail into significant applications that will improve marine trade, by better monitoring and protecting the ships. Ship detection from satellite images using artificial neural networks is focused on in this project. State-of-the-art models, including ANN, CNN, and YOLO-v5 algorithms, are utilized for semantic segmentation, alongside computer vision techniques. The ships are successfully detected, and relatively good segmentation masks are extracted in the end. Key Words: Ship Detection, Satellite Images, Semantic Segmentation.